Surgical Treatment of Spinal Meningiomas

Spinal meningiomas are common spinal tumors; in most cases they are benign and with a good surgical prognosis. However, specific location, infiltration of spinal cord, vascular encasement, or spinal root involvement can bring to a less favorable progno‐ sis. We reviewed a series of 173 consecutive patients with spinal meningiomas treated from 1976 to 2011 in our institution, and data were stratified according to sex, age, symptoms, axial location, Simpson resection grade, and functional pre-/postoperative status. Particular attention was paid to description of those factors leading to a poor outcome. Functional improvement at follow-up was observed in 86.7% of cases, 6.4% of patients resulted stable, and 6.9% worsened; a low functional grade before surgery was connected to a lesser improvement after. Anterolateral meningiomas were the most represented (42.2%); a gross total resection (Simpson grade I and II) was conducted in 98.8% and a macroscopically complete removal without dural resection or coagula‐ tion (Simpson grade III) was performed in 1.2%. According to data from our series, negative prognostic factors seem to be: anterior or anterolateral axial implant, long-lasting symptoms before diagnosis, WHO grade > I, Simpson grade II and III resec‐ tion, sphincter involvement, and worse functional grade at onset. or spinal root involvement can bring on to a less favorable prognosis. We can adfirm that the negative prognostic factors in our study were: anterior or anterolateral axial location, pro‐ tracted symptoms before diagnosis, WHO Grade > I, Simpson grade II and III resection, sphincter involvement, and worse functional grade at onset.


Introduction
Spinal meningiomas (SM) account for 1.2-12% of CNS meningiomas, being relatively uncommon [1][2][3]. The typical clinical presentation consists of pain [4], followed by gait, sensory, and sphincter disturbances. The constant improvement of neuroimaging techniques, the use of intraoperative neuromonitoring, and the increasing reliability of the contemporary surgical tools have further improved the prognosis of SM, already recognized as excellent by Cushing and Eisenhardt [5], more than 80 years ago, in their historical monograph. Since this cornerstone in SM literature, many series confirmed this finding [1,[6][7][8][9]. However, cases of anteriorly located or calcified lesions, cases of recurrences, and cases in which there is violation of the arachnoidal layer, invasion of spinal cord parenchyma, and encasement of vascular structure stand as exceptions to this rule.

Incidence, location, and subtypes
SM count for 25-46% of the intradural-extramedullary lesions, second to spinal schwannomas, which is the most frequent entity of this location [10]. SM favor females [4,6,11,12], its incidence is significantly higher in Caucasians and Asian Pacific islanders than in Afro-Americans and Native Americans [13].
Despite being described at every age [14,15], the highest incidence is in the fifth and sixth decade age group [4,6,7,11]. Below the age of 18, predilection for female sex is not present. This feature, matched with a peak of incidence during peri-menopausal period, is consistent with the widely described and recognized neoplastic cells responsitivity to female sex hormones, similar to what happens for intracranial meningiomas [11,12]. The exact incidence of SM is not known even though is estimated about 0.5-2 cases per 100,000 persons per year [3].
The relatively higher incidence in the thoracic spine is also noted in current literature compared to the other biomechanically active segments [6][7][8]16]. The dural attachment is most frequently found in the anterior/anterolateral dura. Our series (Figure 1 and Table 1) confirms these data [1,2,6,7].  SM histological subtypes are the same as their intracranial counterpart. The percentage of atypical/anaplastic SM is significantly lower than intracranial meningiomas [12]. In our series, WHO Grade I ( Table 1) was the most represented subgroup of lesions in our cohort (98.3% of the lesions). These data are similar to other series [4,6,11,12].
A higher incidence of SM is also well documented in patients suffering from neurofibromatosis. When a SM is detected in early life, it should raise suspicion of neurofibromatosis, especially in patients below the fourth decade and with extrathoracic sagittal location [12].

Symptoms and physical findings
Literature recognize pain as the typical symptom at clinical onset of SM [6,7,17]. Pain can be axial, radicural, or radiating to upper or lower limbs, depending on the location of the tumor; it is commonly associated with paresthesias, hot and cold sensations, and sensory disturbances, followed by gait instability evolving in an obvious motor deficit, which is usually late due to the typical growth slowness of this lesions [2]. Sphincter impairment is a late finding and involves from 15 to 40% of the patients [18].
In our cohort, the most common disturbance at clinical onset was pain in 32.9% of cases (57 patients), followed by motor and gait disorder in 31.8% of cases, (55 patients); less frequently we observed paresthesias (30.6%, 53 patients), and occasionally a pure sensory disturbance was detectable (4.6%, 8 patients). Only 18.5% of patients complained a single disturbance at onset. The second disturbance in order of appearance was a motor disorder (61 patients, 35.3%) and occasionally a sphincter disturbance (10 patients, 5.8%). The average duration of symptoms was 20.01±18.86 months (range, 0-120 months) ( Table 1).

Radiology
Standard X-ray plain film is of limited value, it may demonstrate pedicle or soma erosion, abnormalities in the normal spine curvatures [18]. X-ray standard myelography may outline a contrast block at the level of the extradural lesion, but it has been disused in the common everyday clinical practice. Spine contrast-enhanced CT scans provide detailed information about bony anatomy; it finds a contemporary huge value while investigating extensively calcified lesions or in defining the detailed anatomy of recurrent SM [6,17].
Since its introduction (at our Institution in 1991), a preoperative magnetic resonance imaging (MRI) scans have become the gold standard in defining this pathology. T1w, T2w, T1w gadolinium-enhanced sequences are routinely performed to investigate anatomical details of the lesions (Figure 2A, B, E, and F). Gadolinium-enhanced MRI imaging demonstrated an homogeneously enhanced mass delocating spinal cord parenchyma and nerve roots, intraextradural extension of the lesion, and dural tail; it may not sufficiently help the differential diagnosis between spinal schwannomas and meningiomas [14].

Surgical treatment-technical standards, pitfalls, and advances
Standard prone position is commonly used to approach a SM. Mayfield clamp may prove useful in cases of high subaxial cervical SM. A preoperative radioscopic localization of the involved levels is routinely carried out. After a standard midline skin incision and hypodermal tissues sharp dissection, fascia is incised in a standard midline fashion, and subperiostal exposure of laminae and articular processes is performed. A standard laminotomy (or hemilaminectomy) exposes the dural sac. In case of thoracic spine anterior/anterolateral axially located lesion, a costo-trasversectomy to enlarge surgical corridor can be performed. Midline temporary 4 or 5-0 silk tack-up sutures lead to a safer midline dural opening. Under magnification of the operative microscope, a microsurgical lesionectomy is performed (Figure 2C and D). Regardless of the potential infiltrative pattern or encasement of spinal vascular structures and nerve roots, a CUSA debulking is made to avoid any traction on the spinal cord.
Total removal is the target in all posteriorly located lesions, dural attachment must be removed, and a duraplasty completes the procedure. For anterior/anterolateral lesions, macroscopically complete excision is always associated with generous coagulation of the dural attachment.
Experience and literature have led us to strongly prefer laminotomy over laminectomy to perform a standard posterior midline approach for the evidence-based risk of iatrogenic postoperative instability. In order to clarify the difference between laminectomy and laminotomy, we use the term "laminotomy" to describe a "partial laminectomy" sparing of the medial facet of the articular process.
The reported incidence of postoperative instability in cervical spine is as high as 56% in more than four-level laminectomies and 11% in less than four level [19], whereas in thoracolumbar spine, instability appears in 25% of patients receiving more than two-level laminectomies [20]. Postoperative deformity is reported in 9.4% of patients receiving laminectomy compared to 3% of patients receiving laminotomy [21].
In most cases, a standard posterior approach allows surgeon to work in an adequate surgical corridor to achieve complete excision even for anterior/anterolateral lesions. In our experience, standard posterior approach is the gold standard in huge number of cases. However, intraextradural lesion with infiltration of the vertebral body, massively calcified lesions, anterior/ anterolateral lesions, or recurrent tumors with spinal cord invasion stand as notable exceptions to this rule. In these cases, lateral or anterior approaches, with generous arthrectomy, pedicle resection, and partial/total vertebrectomy, performed to gain an optimal attachment control and safer spinal cord manipulation may be required [8,16,22,23].

Neurological and functional outcome and complications
SM are slowly growing lesion, typically benign. Usually, they carry a fair neurological and functional prognosis [1,[6][7][8][9]. However, cases of anteriorly located or calcified lesions, recurrences, and cases in which there is violation of the arachnoidal layer, invasion of spinal cord parenchyma, and encasement of vascular structure stand as exceptions to this rule.
For this reason, one of the major efforts in SM surgery research and literature has historically been to preoperatively identify cases with a worse functional and neurological prognosis.
With the same aforementioned purpose, we critically reviewed our entire cohort of patients operated on for SM excision through a detailed retrospective analysis. Functional and neurological data about the outcome of each patient were recoded with Frankel [24] and McCormick [25] scales.
Frankel scale is a functional evaluation scale, initially conceived for spinal cord injury (SCI) but capable of assessing residual function "below" the level of a lesion. It is extremely easy The scales are reported in Table 2.  In our sample, a long duration of preoperative symptoms and severe preoperative impairment are related to worse outcomes; this result is similar to other reports [8,26]. The author (A.R.) has previously reported a clear possibility of recovery even in patients harboring deep functional preoperative impairment [26], outlining that age (under 60) and duration of preoperative impairment stand as the most important predictors of neurological recovery.
Anteriorly located lesions imply the most difficult setting in SM surgery [4,7,8,12]. Consistent with the literature [7,8], in our series, anterior/anterolateral lesions were identified as a subgroup with worse outcome compared to different axial topographies. Purely anterior lesions with arachnoidal adhesions to spinal cord and vascular encasement lead to an increased risk of spinal cord traction and surgical damage.
Some authors report massive calcification of the lesion as a predictor of neurological postoperative deterioration [6,8]. In our series, only three patients harbored massively calcified spinal meningiomas, and thus this feature failed to reach statistical significance to confirm these findings.
Recurrence of SM in our sample was predictive of worse outcome. Our recurrence rate was 2.3% (4/173 cases). In the current literature, this rate varies between 1.3 and 13% [2,4,6,9,14,18,27,28]. Obviously, WHO grading of the lesions is correlated with the probability of local relapses. In recurrent SM surgery, arachnoidal scarring caused by the first procedure can make tumor debulking harder and, despite the regular use of intraoperative neuromonitoring and CUSA, can lead to a worse outcome. In our cohort, patients harboring recurrent lesions were 4, a total of just two patients underwent a second SM excision surgery, and, although underrepresented compared to the entire sample (2 of 4 patients on 173 patients of the entire cohort), showed a statistical association with worse outcomes. Therefore, whenever feasible, our experience suggests, according to Literature, that first surgery should always be as aggressive as possible [7,23].
No postoperative death was recorded in our sample. Mortality rate in SM surgery is usually extremely low, ranging between 0.8 and 2% in series reported after 1999 [6][7][8].

Complications
In SM surgery, complications can be coarsely divided in two subgroups: neurological postoperative worsening and surgical complications causing neurological impairment. Among the last subgroup, most common surgical complications include spinal epidural hematoma, CSF leakage with or without deep or superficial infections, syringomyelia, and iatrogenic instability [2,6,7,[19][20][21].
The first subgroup was extensively discussed in the previous paragraph as well as concerns about iatrogenic instability are described in surgical treatment section.
Surgical complications counted for a total of 7/173 cases (4.04% of patients). A total of four patients (two postoperative spinal epidural hematoma, one syringomyelia, and one adverse reaction to dural sealant) required a revision surgery. The remaining underwent conservative treatments (three patients with CSF leakage resolved with lumbar drainage).
Lumbar drainage is reserved for the management of at least 4-5 days of persistent wound CSF leakage. In this series, three patients suffered from wound CSF leakage and were treated conservatively by bedrest and lumbar drainage positioning. These complications were completely resolved in 5-7 days and did not require a surgical revision. Lumbar drainage use for CSF leakages treatment is hugely validated by experience and literature [29,30].
No case of deep or superficial infection was recorded in our series. According to contemporary Literature, at present, deep and superficial infections are exceptional complications [7,17].

Simpson grade of resection and recurrences
In the spinal compartment, recurring lesions are less common than meningiomas rising within the intracranial compartment (10-26%) [4]. Mirimanoff et al. [28] showed that the long-term follow-up of patients operated on for spinal meningioma removal can indicate up to 13% of recurrence at 10 years. However, the main limitation of this study is the small size of the cohort investigated (18 patients). Nakamura et al. [23] reported a recurrence rate as high as 30% at 12-year follow-up for patients who received a Simpson grade II resection. Setzer et al. [16] reported a stratification of recurrence probability by WHO histological grading: 1.4% for grade I, 50% for grade II, and 100% for grade III and IV, in 43.5 ± 24.8 months of follow-up.
Simpson grade I was obtained in 52 patients (30%). In all these cases, dural attachment, mainly posterior or posterolateral, was removed and a duraplasty with eterologous dural patch was performed. The rationale for this procedure is that literature analysis reveals a difference in terms of recurrence probability for Simpson grade I and II [2,18,23], and these data are confirmed in our series. In case of duraplasty, dural sealant was routinely used, since its use is substantially supported by literature [31].
Our routine postoperative SM follow-up protocol consists of a minimum 4 years of clinical and radiological follow-up reserved for patients with Simpson grade of resection I and WHO grade I lesions. Simpson grade of resection > I and WHO grade > I patients underwent a closer clinical and radiological follow-up. Such a follow-up protocol appears justified in relation to the natural history of SM, an extremely slowly growing lesion, which favors fifth/sixth decade. Some recurrences may be ignored because they never reach clinical significance in the lifespan of the patient. Our policy toward recurrent lesions is to reoperate only on those with MRIdemonstrated progressive regrowth tendency, because, according to literature and experience, operating a recurrent SM is often technically challenging [7,8,32,33].

Conclusions
Spinal meningiomas are common primary spinal tumors, in most cases benign and with a good surgical prognosis. However, specific location, infiltration of spinal cord, vascular encasement, or spinal root involvement can bring on to a less favorable prognosis. We can adfirm that the negative prognostic factors in our study were: anterior or anterolateral axial location, protracted symptoms before diagnosis, WHO Grade > I, Simpson grade II and III resection, sphincter involvement, and worse functional grade at onset.